mradermacher/LaMini-GPT-774M-19000-steps-polish-GGUF

Same weights, converted vs Lajonbot/LaMini-GPT-774M-19000-steps-polish. Same weights stored in another dtype (F32/U8 → F16/F32).

Tensors changed: Attention 0/72, Norms 0/146, MLP 0/72, Embeddings 0/2

On the Hub · Downloads on Model Pulse